AI SparkupChatbot Arena 1위 모델, 단 2표로 바뀐다는 MIT 연구 결과
MIT 연구진이 발견한 LLM 랭킹 플랫폼의 충격적 취약성. 57,000표 중 단 2표만 제거해도 1위 모델이 바뀌는 현상과 그 의미를 분석합니다.
WinbuzzerLMArena Gets $100M at $600M Valuation for AI Model Testing
#AI #LMArena #AIFunding #ChatbotArena #AIBenchmarks #UCBerkeley
https://winbuzzer.com/2025/05/21/lmarena-gets-100m-at-600m-valuation-for-ai-model-testing-xcxwbn/
Miguel Afonso Caetano"Measuring progress is fundamental to the advancement of any scientific field. As benchmarks play an increasingly central role, they also grow more susceptible to distortion. Chatbot Arena has emerged as the go-to leaderboard for ranking the most capable AI systems. Yet, in this work we identify systematic issues that have resulted in a distorted playing field. We find that undisclosed private testing practices benefit a handful of providers who are able to test multiple variants before public release and retract scores if desired. We establish that the ability of these providers to choose the best score leads to biased Arena scores due to selective disclosure of performance results. At an extreme, we identify 27 private LLM variants tested by Meta in the lead-up to the Llama-4 release. We also establish that proprietary closed models are sampled at higher rates (number of battles) and have fewer models removed from the arena than open-weight and open-source alternatives. Both these policies lead to large data access asymmetries over time. Providers like Google and OpenAI have received an estimated 19.2% and 20.4% of all data on the arena, respectively. In contrast, a combined 83 open-weight models have only received an estimated 29.7% of the total data. We show that access to Chatbot Arena data yields substantial benefits; even limited additional data can result in relative performance gains of up to 112% on the arena distribution, based on our conservative estimates. Together, these dynamics result in overfitting to Arena-specific dynamics rather than general model quality. The Arena builds on the substantial efforts of both the organizers and an open community that maintains this valuable evaluation platform. We offer actionable recommendations to reform the Chatbot Arena’s evaluation framework and promote fairer, more transparent benchmarking for the field."
https://arxiv.org/abs/2504.20879
#AI #GenerativeAI #LLMs #Chatbots #ChatbotArena #Llama #Meta #OpenSource
The Leaderboard Illusion
Measuring progress is fundamental to the advancement of any scientific field. As benchmarks play an increasingly central role, they also grow more susceptible to distortion. Chatbot Arena has emerged as the go-to leaderboard for ranking the most capable AI systems. Yet, in this work we identify systematic issues that have resulted in a distorted playing field. We find that undisclosed private testing practices benefit a handful of providers who are able to test multiple variants before public release and retract scores if desired. We establish that the ability of these providers to choose the best score leads to biased Arena scores due to selective disclosure of performance results. At an extreme, we identify 27 private LLM variants tested by Meta in the lead-up to the Llama-4 release. We also establish that proprietary closed models are sampled at higher rates (number of battles) and have fewer models removed from the arena than open-weight and open-source alternatives. Both these policies lead to large data access asymmetries over time. Providers like Google and OpenAI have received an estimated 19.2% and 20.4% of all data on the arena, respectively. In contrast, a combined 83 open-weight models have only received an estimated 29.7% of the total data. We show that access to Chatbot Arena data yields substantial benefits; even limited additional data can result in relative performance gains of up to 112% on the arena distribution, based on our conservative estimates. Together, these dynamics result in overfitting to Arena-specific dynamics rather than general model quality. The Arena builds on the substantial efforts of both the organizers and an open community that maintains this valuable evaluation platform. We offer actionable recommendations to reform the Chatbot Arena's evaluation framework and promote fairer, more transparent benchmarking for the field
Experts Challenge Validity and Ethics of Crowdsourced AI Benchmarks Like LMArena (Chatbot Arena)
#AI #AIBenchmarks #AIModels #LMArena #ChatbotArena #AIethics #LLMs #AIEvaluation #Crowdsourcing #GenAI
AI Benchmarking Platform Chatbot Arena Forms New Company, Launches LMArena
#AI #GenAI #LLMs #AIChatbots #LMArena #ChatbotArena #AIBenchmarks #AIModels #AIevaluation
alphareeWow! I didn't really like Gemma 2, but Gemma 3, released today, is awesome. It comes in four sizes, 1b, 4b, 12b and 27b. It's super fast and except for the 1b version it can even handle images.
The 27B version apparently outperforms both DeepSeek v3 and LLaMA3-405 on the ChatbotArena benchmark.
It's also the first small model I've tested that's good at German.
#gemma3 #gemma #gemma2 #google #ai #programming #google #model #local #gemini #multimodal #vision #wow #chatbotarena #german
Moreno Colaiacovo#ChatbotArena Italia è una piattaforma che ha l'obiettivo di comparare e valutare i Large Language Models sulla lingua italiana. 🤖🇮🇹
Se volete partecipare, basta sottoporre un prompt a due modelli #AI scelti a caso dal sistema e votare la migliore. C'è anche la classifica!
Matthias SchüsslerDie #Top5 diese Woche im Blog:
5️⃣ #TheStoryGraph als Alternative zu #Goodreads: https://blog.clickomania.ch/2025/02/18/the-storygraph-book-recommendations-review/
4️⃣ Vergleich von LLMs in der #ChatbotArena: https://blog.clickomania.ch/2025/02/20/lmarena-ai-llm-comparison-platform-review/
3️⃣ Eine Kult-Website! https://blog.clickomania.ch/2025/02/21/wikifeet-com-review-and-critical-acclaim/
2️⃣ Wie #BillGates in der Schweiz lange nicht ernst genommen wurde: https://blog.clickomania.ch/2025/02/19/erste-erwaehnung-bill-gates-in-den-schweizer-medien/
1️⃣ #LeChat von Mistral brilliert beim Karin-Keller-Sutter-Test: https://blog.clickomania.ch/2025/02/17/mistral-le-chat-review/
#clickomaniach
5️⃣ #TheStoryGraph als Alternative zu #Goodreads: https://blog.clickomania.ch/2025/02/18/the-storygraph-book-recommendations-review/
4️⃣ Vergleich von LLMs in der #ChatbotArena: https://blog.clickomania.ch/2025/02/20/lmarena-ai-llm-comparison-platform-review/
3️⃣ Eine Kult-Website! https://blog.clickomania.ch/2025/02/21/wikifeet-com-review-and-critical-acclaim/
2️⃣ Wie #BillGates in der Schweiz lange nicht ernst genommen wurde: https://blog.clickomania.ch/2025/02/19/erste-erwaehnung-bill-gates-in-den-schweizer-medien/
1️⃣ #LeChat von Mistral brilliert beim Karin-Keller-Sutter-Test: https://blog.clickomania.ch/2025/02/17/mistral-le-chat-review/
#clickomaniach
Eine globale Rangliste der KI-Sprachmodelle – und die Möglichkeit, mehrere LLMs blind zu vergleichen: Beides gibt es auf #ChatbotArena. Tipp: Hier war Deepseek zu entdecken, bevor der mediale Hype ausgebrochen ist.
https://blog.clickomania.ch/2025/02/20/lmarena-ai-llm-comparison-platform-review/
#clickomaniach
https://blog.clickomania.ch/2025/02/20/lmarena-ai-llm-comparison-platform-review/
#clickomaniach
CEOTECH.ITOpenAI potenzia GPT-4o: scrittore AI più creativo
#AI #ChatbotArena #ChatGPT #Gemini #GenAI #Google #GPT4o #IntelligenzaArtificiale #LLM #Notizie #Novità #OpenAI #TechNews #Tecnologia
https://www.ceotech.it/openai-potenzia-gpt-4o-scrittore-ai-piu-creativo/
